Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.kernel > #60075 > unrolled thread

Linux 4.14.14 kernel bug on line 152 fs/reiser4/block_alloc.c

Back to article view | Back to linux.debian.kernel

Contents

  Linux 4.14.14 kernel bug on line 152 fs/reiser4/block_alloc.c jose@metztli-it.com (Metztli Information Technology) - 2018-01-26 09:40 +0100
    Re: Linux 4.14.14 kernel bug on line 152 fs/reiser4/block_alloc.c Edward Shishkin <edward.shishkin@gmail.com> - 2018-01-26 15:10 +0100

#60075 — Linux 4.14.14 kernel bug on line 152 fs/reiser4/block_alloc.c

Niltze, Ed-

I had been using Linux kernel 4.14.14 built 'the Debian way' for a couple of days
--doing some light VirtualBox development and successful virtual machine test/creation.

Notwithstanding, today 01-25-2018 experienced black screen kernel bug trace during browser usage.
I wrote down line number of reiser4 file.

FYI: No dataloss but had to reboot Debian AMD64 system due to 'kernel bug' at reference file/line:

invalid opcode: 0000 [#2] SMP PTI

file []-4.14.14/linux/fs/reiser4/block_alloc.c
--------------------------------------------------
[...]
/* Static functions for <reiser4 super block>/<reiser4 context> block counters
   arithmetic. Mostly, they are isolated to not to code same assertions in
   several places. */
static void sub_from_ctx_grabbed(reiser4_context * ctx, __u64 count)
{
        BUG_ON(ctx->grabbed_blocks < count); # kernel bug line 152
        assert("zam-527", ctx->grabbed_blocks >= count);
        ctx->grabbed_blocks -= count;
}
[...]
-----------------------------------------------------

In addition to usual prior mentioned specific Linux 4.14.x series logged entries during kernel builds:

fs/reiser4/carry.o: warning: objtool: fatal_carry_error.isra.5() falls through to next function print_level()
fs/reiser4/txnmgr.o: warning: objtool: commit_txnh()+0x580: stack state mismatch: cfa1=7+120 cfa2=7+136
fs/reiser4/flush.o: warning: objtool: reverse_allocate_parent() falls through to next function squalloc_right_twig_cut.isra.27()
fs/reiser4/wander.o: warning: objtool: reiser4_write_logs()+0x5a2: stack state mismatch: cfa1=7+184 cfa2=7+200
fs/reiser4/vfs_ops.o: warning: objtool: .text: unexpected end of section


Best Professional Regards.

-- 
Jose R R
http://metztli.it
---------------------------------------------------------------------------------------------
Download Metztli Reiser4: Debian Stretch w/ Linux 4.14 AMD64
---------------------------------------------------------------------------------------------
feats ZSTD compression https://sf.net/projects/metztli-reiser4/

[toc] | [next] | [standalone]

#60076

On 01/26/2018 09:15 AM, Metztli Information Technology wrote:
> Niltze, Ed-

Hello.

>
> I had been using Linux kernel 4.14.14 built 'the Debian way' for a couple of days
> --doing some light VirtualBox development and successful virtual machine test/creation.
>
> Notwithstanding, today 01-25-2018 experienced black screen kernel bug trace during browser usage.
> I wrote down line number of reiser4 file.
>
> FYI: No dataloss but had to reboot Debian AMD64 system due to 'kernel bug' at reference file/line:
>
> invalid opcode: 0000 [#2] SMP PTI
>
> file []-4.14.14/linux/fs/reiser4/block_alloc.c
> --------------------------------------------------
> [...]
> /* Static functions for <reiser4 super block>/<reiser4 context> block counters
>     arithmetic. Mostly, they are isolated to not to code same assertions in
>     several places. */
> static void sub_from_ctx_grabbed(reiser4_context * ctx, __u64 count)
> {
>          BUG_ON(ctx->grabbed_blocks < count); # kernel bug line 152
>          assert("zam-527", ctx->grabbed_blocks >= count);
>          ctx->grabbed_blocks -= count;
> }
> [...]

This is an old bug: a leak of grabbed disk space (reservation).
If it is reproducible, then one needs to insert debugging messages
to find out where the grabbed space disappears. Otherwise, there
is no other means except code review.

> -----------------------------------------------------
>
> In addition to usual prior mentioned specific Linux 4.14.x series logged entries during kernel builds:
>
> fs/reiser4/carry.o: warning: objtool: fatal_carry_error.isra.5() falls through to next function print_level()
> fs/reiser4/txnmgr.o: warning: objtool: commit_txnh()+0x580: stack state mismatch: cfa1=7+120 cfa2=7+136
> fs/reiser4/flush.o: warning: objtool: reverse_allocate_parent() falls through to next function squalloc_right_twig_cut.isra.27()
> fs/reiser4/wander.o: warning: objtool: reiser4_write_logs()+0x5a2: stack state mismatch: cfa1=7+184 cfa2=7+200
> fs/reiser4/vfs_ops.o: warning: objtool: .text: unexpected end of section

I am aware of those warnings: a friendship with objtool is required.
Not of high priority.

Thanks,
Edward.

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.kernel

csiph-web